The Scientific Bases
THE SUCCESS OF PERIOPROJECT
More than 65,000 data analysed on 500 patients over 25 years
This predictive model´s database is one of the largest samples to monitor tooth loss due to periodontal disease in patients following periodontal maintenance (Martinez-Canut, P. (2015) Predictors for tooth loss due to periodontal disease in patients following long-term periodontal maintenance. Journal of Clinical Periodontology 42, 1115-1125).
About twenty factors predictive of tooth loss were analysed using different statistical tests, searching for consistency of results. Eleven of these factors, most clearly associated with tooth loss, were finally selected to construct the predictive model.
The basic idea behind a predictive model is to obtain the probability of tooth loss as a value from 0 to 1, rather than using conventional prognostic categories, which are subjectively interpreted according to regression coefficients (e.g., p=0.0005). This probability value of tooth loss can also be associated with ranges of tooth survival expectancy, facilitating decision making.
PerioProject is not a tool capable of assuring which teeth will be lost and when, but it does give extreme values indicating a high probability that the tooth will be maintained (99% specificity) or lost, while estimating the survival expectancy for intermediate values, and thus providing real meaning to the conventional questionable prognosis category.
The following PDF document addresses the scientific bases and the development of the predictive model